Search NASA⌕ Search

SEARCH · Search NASA

Results for “evaluation datasets”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,549 records · Page 86

Breast Imaging Reporting and Data System (BI-RADS) breast composition descriptors: Automated measurement development for full field digital mammography

Purpose: The Breast Imaging Reporting and Data System (BI-RADS) breast composition descriptors are used for standardized mammographic reporting and are assessed visually. This reporting is clinically relevant because breast composition can impact mammographic sensitivity and is a breast cancer risk factor. New techniques are presented and evaluated for generating automated BI-RADS breast composition descriptors using both raw and calibrated full field digital mammography (FFDM) image data.Methods: A matched case-control dataset with FFDM images was used to develop three automated measures for the BI-RADS breast composition descriptors. Histograms of each calibrated mammogram in the percent glandular (pg) representation were processed to create the new BR{sub pg} measure. Two previously validated measures of breast density derived from calibrated and raw mammograms were converted to the new BR{sub vc} and BR{sub vr} measures, respectively. These three measures were compared with the radiologist-reported BI-RADS compositions assessments from the patient records. The authors used two optimization strategies with differential evolution to create these measures: method-1 used breast cancer status; and method-2 matched the reported BI-RADS descriptors. Weighted kappa (κ) analysis was used to assess the agreement between the new measures and the reported measures. Each measure's association with breast cancer was evaluated with odds ratios (ORs) adjusted for body mass index, breast area, and menopausal status. ORs were estimated as per unit increase with 95% confidence intervals.Results: The three BI-RADS measures generated by method-1 had κ between 0.25–0.34. These measures were significantly associated with breast cancer status in the adjusted models: (a) OR = 1.87 (1.34, 2.59) for BR{sub pg}; (b) OR = 1.93 (1.36, 2.74) for BR{sub vc}; and (c) OR = 1.37 (1.05, 1.80) for BR{sub vr}. The measures generated by method-2 had κ between 0.42–0.45. Two of these measures were significantly associated with breast cancer status in the adjusted models: (a) OR = 1.95 (1.24, 3.09) for BR{sub pg}; (b) OR = 1.42 (0.87, 2.32) for BR{sub vc}; and (c) OR = 2.13 (1.22, 3.72) for BR{sub vr}. The radiologist-reported measures from the patient records showed a similar association, OR = 1.49 (0.99, 2.24), although only borderline statistically significant.Conclusions: A general framework was developed and validated for converting calibrated mammograms and continuous measures of breast density to fully automated approximations for the BI-RADS breast composition descriptors. The techniques are general and suitable for a broad range of clinical and research applications.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Construction of Polarimetric Radar-Based Reference Rain Maps for the Iowa Flood Studies Campaign

The Global Precipitation Measurement (GPM) Mission Iowa Flood Studies (IFloodS) campaign was conducted in central and northeastern Iowa during the months of April-June, 2013. Specific science objectives for IFloodS included quantification of uncertainties in satellite and ground-based estimates of precipitation, 4-D characterization of precipitation physical processes and associated parameters (e.g., size distributions, water contents, types, structure etc.), assessment of the impact of precipitation estimation uncertainty and physical processes on hydrologic predictive skill, and refinement of field observations and data analysis approaches as they pertain to future GPM integrated hydrologic validation and related field studies. In addition to field campaign archival of raw and processed satellite data (including precipitation products), key ground-based platforms such as the NASA NPOL S-band and D3R Ka/Ku-band dual-polarimetric radars, University of Iowa X-band dual-polarimetric radars, a large network of paired rain gauge platforms, and a large network of 2D Video and Parsivel disdrometers were deployed. In something of a canonical approach, the radar (NPOL in particular), gauge and disdrometer observational assets were deployed to create a consistent high-quality distributed (time and space sampling) radar-based ground "reference" rainfall dataset, with known uncertainties, that could be used for assessing the satellite-based precipitation products at a range of space/time scales. Subsequently, the impact of uncertainties in the satellite products could be evaluated relative to the ground-benchmark in coupled weather, land-surface and distributed hydrologic modeling frameworks as related to flood prediction. Relative to establishing the ground-based "benchmark", numerous avenues were pursued in the making and verification of IFloodS "reference" dual-polarimetric radar-based rain maps, and this study documents the process and results as they pertain specifically to efforts using the NPOL radar dataset. The initial portions of the "process" involved dual-polarimetric quality control procedures which employed standard phase and correlation-based approaches to removal of clutter and non-meteorological echo. Calculation of a scale-adaptive KDP was accomplished using the method of Wang and Chandrasekar (2009; J. Atmos. Oceanic Tech.). A dual-polarimetric blockage algorithm based on Lang et al. (2009; J. Atmos. Oceanic Tech.) was then implemented to correct radar reflectivity and differential reflectivity at low elevation angles. Next, hydrometeor identification algorithms were run to identify liquid and ice hydrometeors. After the quality control and data preparation steps were completed several different dual-polarimetric rain estimation algorithms were employed to estimate rainfall rates using rainfall scans collected approximately every two to three minutes throughout the campaign. These algorithms included a polarimetrically-tuned Z-R algorithm that adjusts for drop oscillations (via Bringi et al., 2004, J. Atmos. Oceanic Tech.), and several different hybrid polarimetric variable approaches, including one that made use of parameters tuned to IFloodS 2D Video Disdrometer measurements. Finally, a hybrid scan algorithm was designed to merge the rain rate estimates from multiple low level elevation angle scans (where blockages could not be appropriately corrected) in order to create individual low-level rain maps. Individual rain maps at each time step were subsequently accumulated over multiple time scales for comparison to gauge network data. The comparison results and overall error character depended strongly on rain event type, polarimetric estimator applied, and range from the radar. We will present the outcome of these comparisons and their impact on constructing composited "reference" rainfall maps at select time and space scales.

Radar↗

Regional Feedstock Partnership Biomass Quality Assessment Final Report

The United States (U.S.) Department of Energy (DOE) developed the Billion-Ton Vision to enable production of one-billion tons of sustainable, reliable biomass for the bioenergy industry by 2030 (Perlack et al., 2005). The Sun Grant Regional Feedstock Partnership (RFP) was organized to fill information gaps and validate biomass yield assumptions related to the Billion-Ton Study (Owens, 2018; Owens, Karlen, and Lacey, 2016). Along with the more than 130 scientific publications generated from these studies, yield and sustainability data from the RFP field trials not only validated the Billion-Ton estimates, but were critical in developing both the U.S. Billion-Ton Update report in 2011 and the 2016 Billion-Ton Report (DOE, 2011; 2016). The intention of this biomass quality assessment report is to build on these initial successes from the RFP field trials by focusing on variability in biomass quality data necessary to evaluate conversion performance. This report contains a summary of chemical quality results from samples collected as part of the RFP field trials. This report focuses on assessment of the impact of experimental agronomic designs on biomass properties followed by analyses of the impact of environmental and production variables on biomass properties. Datasets include species and other genetic variables, fertilizer treatments, harvest information, and yield, as well as other publicly available data such as precipitation, temperature, soil properties, and drought. The key outcomes from this chemical quality focused assessment have included: • Complete evaluation of the impacts of agronomic designs, genetics, and environmental conditions on chemical properties for Miscanthus, switchgrass, sorghum, energycane, mixed perennial grasses, and shrub willow short-rotation feedstocks • Over 30 peer review publications and technical reports focused on variability in quality data • Development of spatial and temporal environmental quality prediction maps for Miscanthus and switchgrass feedstocks allowing for comprehensive evaluation of variability in feedstock chemical quality across U.S. regions and over multiple harvest years

09 BIOMASS FUELS↗

Evaluation of Cloud-Resolving Model Intercomparison Simulations Using TWP-ICE Observations: Precipitation and Cloud Structure

The Tropical Warm Pool – International Cloud Experiment (TWP-ICE) provided high quality model forcing and observational datasets through which detailed model and observational intercomparisons could be performed. In this first of a two part study, precipitation and cloud structures within nine cloud-resolving model simulations are compared with scanning radar reflectivity and satellite infrared brightness temperature observations during an active monsoon period from 19 to 25 January 2006. Most simulations slightly overestimate volumetric convective rainfall. Overestimation of simulated convective area by 50% or more in several simulations is somewhat offset by underestimation of mean convective rain rates. Stratiform volumetric rainfall is underestimated by 13% to 53% despite overestimation of stratiform area by up to 65% because stratiform rain rates in every simulation are much lower than observed. Although simulations match the peaked convective radar reflectivity distribution at low levels, they do not reproduce the peaked distributions observed above the melting level. Simulated radar reflectivity aloft in convective regions is too high in most simulations. In stratiform regions, there is a large spread in model results with none resembling observed distributions. Above the melting level, observed radar reflectivity decreases more gradually with height than simulated radar reflectivity. A few simulations produce unrealistically uniform and cold 10.8-μm infrared brightness temperatures, but several simulations produce distributions close to observed. Assumed ice particle size distributions appear to play a larger role than ice water contents in producing incorrect simulated radar reflectivity distributions aloft despite substantial differences in mean graupel and snow water contents across models.

Varble, Adam↗

Transfer of Physical and Hydraulic Properties Databases to the Hanford Environmental Information System - PNNL Remediation Decision Support Project, Task 1, Activity 6

This report documents the requirements for transferring physical and hydraulic property data compiled by PNNL into the Hanford Environmental Information System (HEIS). The Remediation Decision Support (RDS) Project is managed by Pacific Northwest National Laboratory (PNNL) to support Hanford Site waste management and remedial action decisions by the U.S. Department of Energy and one of their current site contractors - CH2M-Hill Plateau Remediation Company (CHPRC). The objective of Task 1, Activity 6 of the RDS project is to compile all available physical and hydraulic property data for sediments from the Hanford Site, to port these data into the Hanford Environmental Information System (HEIS), and to make the data web-accessible to anyone on the Hanford Local Area Network via the so-called Virtual Library.1 These physical and hydraulic property data are used to estimate parameters for analytical and numerical flow and transport models that are used for site risk assessments and evaluation of remedial action alternatives. In past years efforts were made by RDS project staff to compile all available physical and hydraulic property data for Hanford sediments and to transfer these data into SoilVision®, a commercial geotechnical software package designed for storing, analyzing, and manipulating soils data. Although SoilVision® has proven to be useful, its access and use restrictions have been recognized as a limitation to the effective use of the physical and hydraulic property databases by the broader group of potential users involved in Hanford waste site issues. In order to make these data more widely available and useable, a decision was made to port them to HEIS and to make them web-accessible via a Virtual Library module. In FY08 the original objectives of this activity on the RDS project were to: (1) ensure traceability and defensibility of all physical and hydraulic property data currently residing in the SoilVision® database maintained by PNNL, (2) transfer the physical and hydraulic property data from the Microsoft Access database files used by SoilVision® into HEIS, which is currently being maintained by CHRPC, (3) develop a Virtual Library module for accessing these data from HEIS, and (4) write a User's Manual for the Virtual Library module. The intent of these activities is to make the available physical and hydraulic property data more readily accessible and useable by technical staff and operable unit managers involved in waste site assessments and remedial action decisions for Hanford. In FY08 communications were established between PNNL and staff from Fluor-Hanford Co. (who formerly managed HEIS) to outline the design of a Virtual Library module that could be used to access the physical and hydraulic property data that are to be transferred into HEIS. Data dictionaries used by SoilVision® were also provided to Fluor-Hanford personnel (who are now with CHPRC). During ongoing work to ensure traceability and defensibility of all physical and hydraulic property data that currently reside in the SoilVision® database, it was recognized that further work would be required in this effort before the data were actually ported into HEIS. Therefore work on the Virtual Library module development and an accompanying User's Guide was deferred until an unspecified later date. In FY09 efforts have continued to verify the traceability and defensibility of the physical and hydraulic property datasets that are currently being maintained by PNNL. Although this is a work in progress, several of these datasets should be ready for transfer to HEIS in the very near future. This document outlines a plan for the migration of these datasets into HEIS.

54 ENVIRONMENTAL SCIENCES↗

Coatomer subunit beta 2 (COPB2), identified by label-free quantitative proteomics, regulates cell proliferation and apoptosis in human prostate carcinoma cells

Highlights: • Our paper offered an available method to identified and confirm some differential expression of genes in prostate cancer. • Our paper provided us a common method to research the function of gene for prostate cancer. • Our paper helps us to identify a novel cancer related gene to explain for development of novel targeted drugs. Label-free quantitative proteomics has broad applications in the identification of differentially expressed proteins. Here, we applied this method to identify differentially expressed proteins (such as coatomer subunit beta 2 [COPB2]) and evaluated the functions and molecular mechanisms of these proteins in prostate cancer (PCA) cell proliferation. Proteins extracted from surgically resected PCA tissues and adjacent tissues of 3 patients were analyzed by label-free quantitative proteomics. The target protein was confirmed by bioinformatics and GEO dataset analyses. To investigate the role of the target protein in PCA, we used lentivirus-mediated small-interfering RNA (siRNA) to knockdown protein expression in the prostate carcinoma cell line, CWR22RV1 cells and assessed gene and protein expression by reverse transcription quantitative polymerase chain reaction and western blotting. CCK8 and colony formation assays were conducted to evaluate cell proliferation. Cell cycle distributions and apoptosis were assayed by flow cytometry. We selected the differentiation-related protein COPB2 as our target protein based on the results of label-free quantitative proteomics. High expression of COPB2 was found in PCA tissue and was related to poor overall survival based on a public dataset. Cell proliferation was significantly inhibited in COPB2-knockdown CWR22RV1 cells, as demonstrated by CCK8 and colony formation assays. Additionally, the apoptosis rate and percentage of cells in the G{sub 1} phase were increased in COPB2-knockdown cells compared with those in control cells. CDK2, CDK4, and cyclin D1 were downregulated, whereas p21 Waf1/Cip1 and p27 Kip1 were upregulated, affecting the cell cycle signaling pathway. COPB2 significantly promoted CWR22RV1 cell proliferation through the cell cycle signaling pathway. Thus, silencing of COPB2 may have therapeutic applications in PCA.

60 APPLIED LIFE SCIENCES↗

Stream-reach Identification for New Run-of-River Hydropower Development through a Merit Matrix Based Geospatial Algorithm

Even after a century of development, the total hydropower potential from undeveloped rivers is still considered to be abundant in the United States. However, unlike evaluating hydropower potential at existing hydropower plants or non-powered dams, locating a feasible new hydropower plant involves many unknowns, and hence the total undeveloped potential is harder to quantify. In light of the rapid development of multiple national geospatial datasets for topography, hydrology, and environmental characteristics, a merit matrix based geospatial algorithm is proposed to help identify possible hydropower stream-reaches for future development. These hydropower stream-reaches sections of natural streams with suitable head, flow, and slope for possible future development are identified and compared using three different scenarios. A case study was conducted in the Alabama-Coosa-Tallapoosa (ACT) and Apalachicola-Chattahoochee-Flint (ACF) hydrologic subregions. It was found that a merit matrix based algorithm, which is based on the product of hydraulic head, annual mean flow, and average channel slope, can help effectively identify stream-reaches with high power density and small surface inundation. The identified stream-reaches can then be efficiently evaluated for their potential environmental impact, land development cost, and other competing water usage in detailed feasibility studies . Given that the selected datasets are available nationally (at least within the conterminous US), the proposed methodology will have wide applicability across the country.

Pasha, M. Fayzul K.↗

Power System Waveform Datasets for Machine Learning

The desire for increased visibility across the electricity grid will necessarily increase the deployment of sensing and measurement devices and associated data management needs to unprecedented levels. For the existing sensing and measurement infrastructure, there remains a great amount of “value” yet to be extracted through advanced data management and analytics. Availability of more data will not, by itself, lead to changes in grid visibility, security, and resiliency. To create the predictive and prescriptive environment required to enable new markets and transactions for customer revenue and a reliable grid, the data must be collected, organized, evaluated, and analyzed using sophisticated algorithms to provide actionable information allowing operators and customers to reliably manage an increasingly complex grid. Progress in artificial intelligence (AI) has been largely driven by large, publicly available datasets that can be used to train AI algorithms such as MNIST, a database of handwritten images of digits, and ImageNet, an image database of everyday objects. These types of publicly available databases of real-world training datasets have been largely credited for advancement of image processing, computer vision, and deep learning algorithms that these use cases deploy. However, in the power systems industry to date, there are few databases with proper event labeling, and data access to a publicly available collection of power system event waveforms that will allow users to interact with grid signature data. Publicly available datasets of power system event waveforms, such as the DOE/EPRI dataset, often lack critical metadata or contain limited examples of each event type, and data formats vary widely across these datasets.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Automated detection of part quality during two-photon lithography via deep learning

Two-photon lithography (TPL) is an additive manufacturing technique for fabricating three-dimensional objects with nanoscale features. A main challenge of TPL is the routine and labor-intensive task of finding suitable light dosage parameters, i.e. writing speed and laser intensity that induce photo-polymerization within a wide variety of candidate photo-curing polymers. Another challenge is the monitoring required during fabrication. In this work, we apply machine learning (ML) models to accelerate the process of identifying optimal light dosage parameters and automate the detection of part quality. We curate TPL videos of different parts fabricated under a range of light dosage parameters using different resins and train spatial-temporal ML models on this data. Our results show that ML models can detect TPL part quality with a 95.1% accuracy in milliseconds. We also evaluate classification failures and identify two operating modes: parameter optimization and part quality detection. Last but not least, we publicly release this labelled dataset so that it may serve as a useful benchmark to the community. Our approach to process optimization and part quality detection addresses important aspects of TPL industrialization, is applicable beyond TPL and should benefit other additive manufacturing techniques with similar barriers to operating at industrial scale.

36 MATERIALS SCIENCE↗

A data-driven CO 2 leakage detection using seismic data and spatial–temporal densely connected convolutional neural networks

In carbon capture and sequestration (also known as carbon capture and storage, or CCS), developing effective monitoring methods is needed to detect and respond to CO 2 leakage. CO 2 leakage detection methods rely on geophysical observations and monitoring sensor network. However, traditional methods usually require the development of site-specific physical models and expert interpretation, and the effectiveness of these methods can be limited to different application locations, operational scenarios, and conditions. Here, we developed a novel data-driven leakage detection method based on densely connected convolutional neural networks. Our method is an end-to-end detection approach, that differs from conventional leakage monitoring methods by directly learning a mapping relationship between seismic data and the CO 2 leakage mass. To account for the spatial and temporal characteristics of seismic data, our novel networks architecture combines 1D and 2D convolutional neural networks. To overcome the computational expense of solving optimization problems, we apply a densely-connecting strategy in our network architecture that reduces the number of network parameters. Based on the features generated by our convolutional neural networks, we further incorporate a long short-term memory network to utilize time-sequential information, which further improves the detection accuracy. Lastly, we employ our detection method to synthetic seismic datasets generated based on flow simulations of a hypothetical CO 2 storage scenario with injection into a partially compartmentalized sandstone storage reservoir. To evaluate method performance, we conducted multiple experiments including a random leakage test, a sequential test, and a robustness test. Numerical results show that our CO 2 leakage detection method successfully detects the leakage and accurately predicts the leakage mass, suggesting that it has the potential for application in monitoring of real CO 2 storage sites.

58 GEOSCIENCES↗

Nuclear Data Sheets for A=154

The experimental results published before Aug 2022 from the various reaction and decay studies leading to nuclides of Z=56 to Z=72, 154 Ba, 154 La, 154 Ce, 154 Pr, 154 Nd, 154 Pm, 154 Sm, 154 Eu, 154 Gd, 154 Tb, 154 Dy, 154 Ho, 154 Er, 154 Tm, 154 Yb, 154 Lu, 154 Hf, in the A=154 mass chain have been reviewed. These data are collected and presented in decay or reaction datasets, together with Adopted Levels and gammas datasets that are the most extensive collections of nuclear structure data for each nuclide. Furthermore this work is intended to supersede the previous evaluation of the A=154 nuclides by C.W. Reich (2009Re14), which was published in Nuclear Data Sheets 110, 2257 (2009).

Nica, N. [Texas A&M University, College Station, T↗

Datasets of Faults in Variable Air Volume Terminal Units in a Multi-Zone Commercial Building

Faults in HVAC systems can decrease system efficiency and equipment lifespan, leading to 5%–30% of energy consumption being wasted in commercial buildings. We identified two common faults in HVAC variable air volume systems: a stuck damper fault in the variable air volume terminal unit and a discharge airflow sensor fault. We conducted three sets of damper stuck tests and two sets of airflow sensor tests, each including a fault-free scenario and scenarios with varying levels of faults, over one day. The faults were implemented in Oak Ridge National Laboratory’s two-story Flexible Research Platform building to generate a high-quality, well-controlled dataset covering fault-induced and fault-free scenarios. The test building, fault test scenarios, and data validation are described here. The open-source dataset includes 1 min intervals of weather and building data on the presence and absence of building faults. This dataset can be used to analyze the effects of HVAC system faults on system operation and indoor building conditions, and to develop or evaluate a fault detection and diagnosis algorithm.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Machine learning approaches for influenza A virus risk assessment identifies predictive correlates using ferret model in vivo data

In vivo assessments of influenza A virus (IAV) pathogenicity and transmissibility in ferrets represent a crucial component of many pandemic risk assessment rubrics, but few systematic efforts to identify which data from in vivo experimentation are most useful for predicting pathogenesis and transmission outcomes have been conducted. To this aim, we aggregated viral and molecular data from 125 contemporary IAV (H1, H2, H3, H5, H7, and H9 subtypes) evaluated in ferrets under a consistent protocol. Three overarching predictive classification outcomes (lethality, morbidity, transmissibility) were constructed using machine learning (ML) techniques, employing datasets emphasizing virological and clinical parameters from inoculated ferrets, limited to viral sequence-based information, or combining both data types. Among 11 different ML algorithms tested and assessed, gradient boosting machines and random forest algorithms yielded the highest performance, with models for lethality and transmission consistently better performing than models predicting morbidity. Comparisons of feature selection among models was performed, and highest performing models were validated with results from external risk assessment studies. Our findings show that ML algorithms can be used to summarize complex in vivo experimental work into succinct summaries that inform and enhance risk assessment criteria for pandemic preparedness that take in vivo data into account.

59 BASIC BIOLOGICAL SCIENCES↗

Assessing the potential of deep learning for protein–ligand docking

The effects of ligand binding on protein structures and their in vivo functions carry numerous implications for modern biomedical research and biotechnology development efforts such as drug discovery. Although several deep learning (DL) methods and benchmarks designed for protein–ligand docking have recently been introduced, so far no previous works have systematically studied the behaviour of the latest docking and structure prediction methods within the broadly applicable context of: (1) using predicted (apo) protein structures for docking (for example, for applicability to new proteins); (2) binding multiple (cofactor) ligands concurrently to a given target protein (for example, for enzyme design); and (3) having no previous knowledge of binding pockets (for example, for generalization to unknown pockets). To enable a deeper understanding of the real-world utility of docking methods, we introduce PoseBench, a comprehensive benchmark for broadly applicable protein–ligand docking. PoseBench enables researchers to rigorously and systematically evaluate DL methods for apo-to-holo protein–ligand docking and protein–ligand structure prediction using both primary ligand and multiligand benchmark datasets, the latter of which we introduce to the DL community. Empirically, using PoseBench, we find that: (1) DL cofolding methods generally outperform comparable conventional and DL docking baseline algorithms, but popular methods such as AlphaFold 3 are still challenged by prediction targets with new protein–ligand binding poses; (2) certain DL cofolding methods are highly sensitive to their input multiple sequence alignments, whereas others are not; and (3) DL methods struggle to strike a balance between structural accuracy and chemical specificity when predicting new or multiligand protein targets.

Morehead, Alex [Lawrence Berkeley National Laborat↗

A machine learning estimator trained on synthetic data for real-time earthquake ground-shaking predictions in Southern California

Abstract After large-magnitude earthquakes, a crucial task for impact assessment is to rapidly and accurately estimate the ground shaking in the affected region. To satisfy real-time constraints, intensity measures are traditionally evaluated with empirical Ground Motion Models that can drastically limit the accuracy of the estimated values. As an alternative, here we present Machine Learning strategies trained on physics-based simulations that require similar evaluation times. We trained and validated the proposed Machine Learning-based Estimator for ground shaking maps with one of the largest existing datasets (<100M simulated seismograms) from CyberShake developed by the Southern California Earthquake Center covering the Los Angeles basin. For a well-tailored synthetic database, our predictions outperform empirical Ground Motion Models provided that the events considered are compatible with the training data. Using the proposed strategy we show significant error reductions not only for synthetic, but also for five real historical earthquakes, relative to empirical Ground Motion Models.

Environmental Sciences & Ecology↗

Application of machine learning for the estimation of electron energy distribution from optical emission spectra

Abstract This paper discusses the use of probabilistic deep neural networks for the prediction of the electron energy probability function in low-temperature non-thermal plasmas. The neural networks are trained using optical emission spectroscopy and Langmuir probe measurements, with the goal of providing a reliable estimate of the electron energy probability function solely from optical emission data. The performance of both non-Bayesian and Bayesian networks is evaluated. It is found that Bayesian models are preferable as they assign a higher level of uncertainty to their prediction especially when the dataset used to train them is small. This work describes one of the many potential applications of machine learning in plasma science and technology.

Physics↗

Fast physics-based launcher optimization for electron cyclotron current drive

With the increased urgency to design fusion pilot plants, fast optimization of electron cyclotron current drive (ECCD) launchers is paramount. Traditionally, this is done by coarsely sampling the 4D parameter space of possible launch conditions consisting of (1) the launch location (constrained to lie along the reactor vessel), (2) the launch frequency, (3) the toroidal launch angle, and (4) the poloidal launch angle. For each initial condition, a ray-tracing simulation is performed to evaluate the ECCD efficiency. Unfortunately, this approach often requires a large number of simulations (sometimes millions in extreme cases) to build up a dataset that adequately covers the plasma volume, which must then be repeated every time the design point changes. Here we adopt a different approach. Rather than launching rays from the plasma periphery and hoping for the best, we instead directly reconstruct the optimal ray for driving current at a given flux surface using a reduced physics model coupled with a commercial ray-tracing code. Repeating this throughout the plasma volume requires only hundreds of simulations, constituting a significant speedup. The new method is validated on two separate example tokamak profiles, and is shown to reliably drive localized current at the specified flux surface with the same optimal efficiency as obtained from the traditional approach.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Designing Next Generation Massively Multithreaded Architectures for Irregular Applications

Irregular applications, such as data mining or graph-based computations, show unpredictable memory/network access patterns and control structures. Massively multi-threaded architectures with large node count, like the Cray XMT, have been shown to address their requirements better than commodity clusters. In this paper we present the approaches that we are currently pursuing to design future generations of these architectures. First, we introduce the Cray XMT and compare it to other multithreaded architectures. We then propose an evolution of the architecture, integrating multiple cores per node and next generation network interconnect. We advocate the use of hardware support for remote memory reference aggregation to optimize network utilization. For this evaluation we developed a highly parallel, custom simulation infrastructure for multi-threaded systems. Our simulator executes unmodified XMT binaries with very large datasets, capturing effects due to contention and hot-spotting, while predicting execution times with greater than 90% accuracy. We also discuss the FPGA prototyping approach that we are employing to study efficient support for irregular applications in next generation manycore processors.

Tumeo, Antonino↗